Which type of regression model do I need and what do I need to do to my variable to allow it to work?

I am trying to build a regression model for price sensitivity. I want to be able to show the number of products ordered based on changes to the list price and out of pocket price for a given product.

After getting a base level assessment, we tested 8 scenarios of price changes, each of which is represented by its own variable. Currently there are 9 independent variables influencing the number of products ordered (my dependent variable). They are as follows:

base List Price = 2,500; Out of Pocket Price(OOP) = 250 (This is the current pricing scenario today and used as the baseline)
scenario_a List Price = 1,250; OOP = 250
scenario_b List Price = 750; OOP = 250
scenario_c List Price = 2,500; OOP = 125
scenario_d List Price = 2,500; OOP = 75
scenario_e List Price = 1,250; OOP = 125
scenario_f List Price = 750; OOP = 75
scenario_g List Price = 1,250; OOP = 375
scenario_h List Price = 3,750; OOP = \$125

I originally thought I would do a multiple non-linear regression model however I realized I am oversimplifying the problem in my head.

Firstly, I'm confused as to how to handle the 2 price components of each variable (scenario). I thought about splitting these apart but since the number of products ordered is dependent on both of these values in each of the scenarios I feel like that would really be the wrong way to go.

I also considered grouping the different scenarios with common List Prices (group 1 - a,e,g; group 2 - b & f... etc) and conducting multiple regressions based on those but I'm not sure if that even makes sense.

Long story short, at least to start out, I need to figure out the appropriate type of regression to run and how, if at all, I need to manipulate / handle the pricing scenario variables to account for the 2 different pricing components.

Also if my syntax is messed up I will do what I need to to fix it. Hopefully it doesn't need to be changed at all but looking at the preview, I'm anticipating the list will get messed up.

• list price, Out Of Pocket Price---can you explain? – kjetil b halvorsen Apr 22 '15 at 14:23
• Each scenario has 2 price components. Think of List Price as MSRP, and OOP as what someone will actually pay – laxpro2001 Apr 22 '15 at 15:46
• and MSRP=Manufacturers Suggested Retail Price? – kjetil b halvorsen Apr 22 '15 at 16:03
• Sorry for not clarifying that. Yes MSRP would be manufacturers suggested retail price or the sticker price. Some people may pay that and some may not it all depends on how the respondent interprets each of the components. – laxpro2001 Apr 22 '15 at 16:25
• Well I learned that I can stack the data-set like I would with a normal linear regression. The next step is to figure out what type of regression to run since I don't think linear makes sense... – laxpro2001 Apr 22 '15 at 23:01